. . . .
. . . .
Caveolin-1 opens endothelial cell junctions
by targeting catenins
Romy Kronstein
1, Jochen Seebach
2, Sylvia Großklaus
1, Carsten Minten
3,
Britta Engelhardt
3, Marek Drab
4, Stefan Liebner
5, Yvan Arsenijevic
6,
Abdallah Abu Taha
2, Tatiana Afanasieva
1, and Hans-Joachim Schnittler
2*
1
Institute of Physiology, Medical Faculty of the TU-Dresden, Dresden, Germany;2
Institute of Anatomy and Vascular Biology, Vesaliusweg 2-4, Westfa¨lische Wilhelms-University of Mu¨nster,
Mu¨nster, Germany;3
Theodor Kocher Institute, University of Bern, Bern, Switzerland;4
Institute of Immunology and Experimental Therapy, Wroclaw, Poland;5
Institute of Neurology, Johann
Wolfgang Goethe-University Frankfurt, Frankfurt, Germany; and6
Jules Gonin Eye Hospital, University of Lausanne, Lausanne, Switzerland Received 8 March 2011; revised 21 September 2011; accepted 26 September 2011; online publish-ahead-of-print 29 September 2011 Time for primary review: 30 days
Aims A fundamental phenomenon in inflammation is the loss of endothelial barrier function, in which the opening of endo-thelial cell junctions plays a central role. However, the molecular mechanisms that ultimately open the cell junctions are largely unknown.
Methods and results
Impedance spectroscopy, biochemistry, and morphology were used to investigate the role of caveolin-1 in the regu-lation of thrombin-induced opening of cell junctions in cultured human and mouse endothelial cells. Here, we demon-strate that the vascular endothelial (VE) cadherin/catenin complex targets caveolin-1 to endothelial cell junctions. Association of caveolin-1 with VE-cadherin/catenin complexes is essential for the barrier function decrease in response to the pro-inflammatory mediator thrombin, which causes a reorganization of the complex in a rope ladder-like pattern accompanied by a loss of junction-associated actin filaments. Mechanistically, we show that in response to thrombin stimulation the protease-activated receptor 1 (PAR-1) causes phosphorylation of caveolin-1, which increasingly associates with b- and g-catenin. Consequently, the association of b- and g-catenin with VE-cadherin is weakened, thus allowing junction reorganization and a decrease in barrier function. Thrombin-induced opening of cell junctions is lost in caveolin-1-knockout endothelial cells and after expression of a Y/F-caveolin-1 mutant but is completely reconstituted after expression of wild-type caveolin-1.
Conclusion Our results highlight the pivotal role of caveolin-1 in VE-cadherin-mediated cell adhesion via catenins and, in turn, in barrier function regulation.
-Keywords b-catenin † Caveolin-1 † Impedance spectroscopy † VE-cadherin † Permeability
1. Introduction
Vascular endothelial (VE) cells form a barrier that controls the exchange of fluid and solutes between the vasculature and the inter-stitial tissue. The calcium-dependent VE-cadherin, and its associated catenins, control cell – cell adhesion, and paracellular barrier function and are important for the formation of tight junction complexes.1,2 The VE-cadherin/catenin complex can be found in all endothelial cells and represents the predominant structure in endothelial junc-tions of postcapillary venules, a location where cell juncjunc-tions prefer-entially open during inflammation and leucocyte diapedesis.1,2 VE-cadherin has a carboxy-terminal domain that binds b-catenin,
g-catenin and p120ctn.1,2 Both b- and g-catenins bind a-catenin,
which was previously suggested to connect the complex directly to actin filaments.1,2 However, this concept was challenged by data
showing that a-catenin is an allosteric molecule incapable of binding to both b-catenin and actin filaments at the same time.3
Recently, the epithelial protein lost in neoplasms (EPLIN) was shown to bind both actin filaments and a-catenin in epithelial cells,4 indicating that the a-catenin/EPLIN protein-chain might
bridge the gap between cadherins and actin filaments. We were able to confirm that endothelial cells of human and mouse origin also express ELPIN (Taha and Schnittler, unpublished data). However, it is generally accepted that the VE-cadherin/catenin
*Corresponding author. Tel:+49 (251) 83 52372; fax: +49 (251) 83 55241, Email: hans.schnittler@uni-muenster.de
Published on behalf of the European Society of Cardiology. All rights reserved.&The Author 2011. For permissions please email: journals.permissions@oup.com.
Cardiovascular Research (2012) 93, 130–140 doi:10.1093/cvr/cvr256
complex is functionally linked to actin filaments1,2and that this inter-action is critical in junction regulation.
Transient increases in paracellular permeability in response to acti-vation of PAR-1 (proteinase-activated receptor-1) by thrombin cause redistribution of VE-cadherin and catenins from a continuous band into a discontinous streak-like pattern with loss of junction-associated actin filaments and stress fibre formation. This process involves the activation of Rho-GTPases, the kinases PI3K, src, and PKC, actin/ myosin contractility, myosin light chain kinases and phosphatases, and the tyrosine phosphatase VEPTP (for review see).1,2In contrast, tightening of cell junctions depends on VE-cadherin clustering accompanied by recruitment of junction-associated actin filaments via rac-1.5 Although a number of relevant mechanisms have been identified, the causal mechanism that ultimately induces opening of endothelial cell junctions in inflammation (e.g. after thrombin admin-istration) is still unknown.
Many studies have provided compelling evidence that caveolin-1, the main scaffolding protein of caveolae,6 binds to signalling mol-ecules7and is involved in regulating endothelial permeability, angio-genesis,8–10 and leucocyte diapedesis.11 Caveolin-1 is composed of a lipophilic, hairpin-shaped, helical sequence embedded in the inner leaflet of the plasma membrane with both N- and C-terminal cyto-plasmic domains. The N-terminus binds to many signalling molecules, including endothelial nitric oxide synthase (eNOS), src kinase, PKCa, and G-proteins and is required for caveolin multimerization.7Previous studies on the role of caveolin-1 in endothelial permeability have reported both positive10,12,13and negative8,14effects. Caveolin-1 can be precipitated with VE-cadherin and E-cadherin13,15 and interacts with b-catenin in zebrafish.16 Thus, caveolin-1 might have a more general role in regulating cell junctions, but its molecular regulation of endothelial cell adhesion and barrier function needs to be defined. Here, we show that localization of caveolin-1 to endothelial cell junctions depends on the VE-cadherin/catenin complexes and that thrombin-induced increase in paracellular permeability is phosphocaveolin-1 dependent.
2. Methods
2.1 Antibodies and reagents
The antibodies and reagents used and their sources are listed in Supplementary material online, (SI-6).
2.2 Primary endothelial cell cultures, Chinese
hamster ovary cells, endothelioma cell lines,
immunofluorescence microscopy, and
transendothelial electrical resistance
Human umbilical cord vein endothelial cells (HUVEC) were isolated from anonymized donors as described17according to the principles outlined in the Declaration of Helsinki; this was approved by the ethics boards of TU-Dresen and the WW-University of Mu¨nster (EK 203112005 and 2009 – 537-f-S, respectively). Cells of passage 1 were used for exper-iments. Wild-type CHO (Chinese hamster ovary) cells and VE-cadherin-expressing CHO cells were cultured as described else-where.18 Caveolin-1-deficient and wild-type endothelioma cell lines were established from hearts of caveolin-1-deficient newborn mice19 and wild-type mice, respectively, as previously described.20 Animals were sacrificed by cervical dislocation in accordance to the German animal protection law and to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication
No.85 – 23, revised 1996). A b-catenin-deficient cell line was a kind gift from Elisabetta Dejana, Milano, and cultured as described.21
Immunolabel-ling was carried out as previously described17 and detected using a laser-scanning microscope (LSM 510 Zeiss, Jena, Germany). Transen-dothelial electrical resistance (TER) of enTransen-dothelial cell cultures was deter-mined from impedance spectroscopy data.5Endothelioma cell cultures
having a TER .3 Vcm2, a value sufficient to evaluate endothelial barrier function by impedance spectroscopy, were used for experiments.
2.3 Generation of recombinant
glutathione-S-transferase-tagged b-catenin
and VE-cadherin cytoplasmic domain
Recombinant murine glutathione-S-transferase (GST)2b-catenin was expressed in E. coli BL21 using the pGEX 2T-1 expression vector (kindly provided by William Weis and S. Pokutta, Stanford, USA) and pur-ified as previously described.22The recombinant cytoplasmic domain of GST-VE-cadherin (GST-VE-cadcyto) was expressed in E. coli strain BL21
using the pGEX 4T-1 expression vector (kindly provided by D. Vestweber, Mu¨nster, Germany). GST-VE-cadcytowas purified using a
GSTrapTM FF glutathione-sepharose column (GE-Healthcare, Uppsala, Sweden) according to the manufacturers’ instruction manual.
2.4 GST pull-down assays
Either GST-tagged b-catenin or VE-cadcyto(65 nM) was bound for 1 h at
48C to glutathione sepharose in binding buffer (20 mM Tris/HCl, pH 7.4, 150 mM NaCl, 1% Triton-X-100) and washed three times in 1 mL of phosphate-buffered saline containing 0.5% Triton X-100, 1 mM dithio-threitol (DTT) and 1 mM PMSF (washing buffer). HUVEC cell cultures (7× 106cells) were lysed for 15 min at 48C in lysis buffer (50 mM Tris/
HCl, pH 7.6, 150 mM NaCl, 1% Triton X-100, 1 mM PMSF, 5 mM DTT, and proteinase inhibitors), and pelleted at 10 000 g for 5 min. Precleared supernatants were subsequently adjusted to 650 mg total protein/mL and a total volume of 800 mL. GST-b-catenin- or GST-VE-cadcyto-loaded
glu-tathione beads were applied to HUVEC cell lysates (7× 106cells) for 2 h at 48C. After three washes using washing buffer, the samples were lysed in SDS-sample buffer and analysed by SDS – PAGE and western blotting.
2.5 DNA plasmids, constructs, and production
of lentiviral vectors
Mouse caveolin-123was cloned into the lentiviral vector pFUWG via the BamHI and EcoRI cloning sites following PCR amplification with forward primer 1 (5′-GATGGATCCATGTCTGGGGGCAAATACGTG-3′, BamHI) and reverse primer 2 (5′-GCCGAATTCTCATATCTCTTTCTGCGTGC-3′, EcoRI). Cav-1-Y14F was generated by PCR amplification of fragment 1 using forward primer 3 (5′-GGACATCTCTTCACTGTTCCC-3′; base exchange A/T is shown in bold) and reverse primer 2, and fragment 2 using forward primer 1 and reverse primer 4 (5′-GGGAA CAGTGAAGAGATGTCC-3′). The two fragments were gel-purified and annealed, and PCR was performed using primers 1 and 2. Purified PCR pro-ducts were digested using BamHI and EcoRI and then cloned into the lenti-viral vector pFUGW using the same cloning sites. Full-length mouse b-catenin24 was amplified by PCR from the pCS2+ vector using the forward primer (5′-CGGCTAGCCACCATGGCTACTCAAGCTGA CCTG-3′) and the reverse primer (5′-CTCTAGAGTTACAGGTC AGTATCAAACCA-3′). The purified PCR product was digested by NheI and XbaI and cloned into the lentiviral vector pFUGW. All constructs were verified by sequencing. Lentiviruses were produced by triple transfec-tion of 293 T cells with pFUGW carrying the gene of interest, packaging vectors, and pMD2G carrying the VSV glycoprotein as described else-where.25Viral stocks were titrated by HIV-1 ELISA according to the manu-facturers’ directions (Zeptometrix Coorperation, Buffalo, USA). Gene transduction into endothelioma cells was performed using 10 TU/cell. Gene transduction efficiency was .90%, as scored by the number of
caveolin-1-positive cells. Lentivirus-mediated gene transfer was non-toxic as tested by determination of endothelial barrier function, growth, and mor-phology. Protein expression was verified by quantitative western blot analy-sis and immunofluorescence microscopy, as indicated.
2.6 Analysis of tyrosine phosphorylation,
immunoprecipitation, and western blotting
Tyrosine phosphorylation assay and quantitation was performed as described elsewhere5 and outlined in the Supplementary materialonline, (SI-7).
2.7 Quantification of western blots, cell
junction proteins, and statistical analyses
Quantitative western blots using total cell extracts (Figure1) were per-formed under standard conditions; prior to gel loading, total protein was determined in SDS-dissolved samples,26 and defined amounts of protein were loaded on each lane. Confluent cultures of HUVEC were pooled (125 cm2 total) and subsequently divided in five different samples, each used for quantitative immunoprecipitation (IP) of VE-cadherin, a-catenin, b-catenin, ZO-1, and PECAM-1 followed by western blotting. The optical density of the antibody band for each IP was set to 1 and co-precipitated protein bands were related to this amount. The same procedure was performed for g-catenin and caveolin-1 (not shown). Relative amounts of co-precipitated proteins shown in Figures4and6were determined by relating the respective protein of stan-dardized IP (compare IP) to the total amount of tubulin obtained fromeach sample. The relative amounts of co-precipitated proteins were obtained by ratio determination as indicated in each figure. Image J-Software (NIH) was used to determine optical density of the specific bands according to instructions.
Cell staining and image acquisitions were performed under standardized conditions. Cell borders were manually selected as described elsewhere5 and fluorescence intensity was determined for each channel separately. All data were tested for significance and standard deviation using GraphPad Software Inc. (San Diego, USA; http://www.graphpad.com).
3. Results
3.1 Caveolin-1 associates with the
VE-cadherin/catenin complex
We first determined that caveolin-1 is an integral component of the VE-cadherin/catenin complex in HUVECs, as indicated by IP of VE-cadherin, a-, b-, or g-catenin, which yielded varying amounts of caveolin-1 and associated junction proteins (Figure 1A and B). IP of b-catenin and its homologue g-catenin yielded the highest amount of caveolin-1 (10% of the total), whereas IP of VE-cadherin yielded much less (Figure1A and B). These interactions were specific as IP of neither zonula occludens protein-1 nor mouse IgG2a co-precipitated caveolin-1, VE-cadherin, or catenins (Figure 1A and B). These results prompted us to test the ability Figure 1 Caveolin-1 is found in a complex with junction molecules in endothelial cells. (A and B) Densitometry-based quantitative western blot analyses of IP of VE-cadherin, a-catenin, b-catenin, ZO-1, g-catenin, PECAM-1, or unspecific mouse IgG2a from same amounts and volume of
Triton X-100 soluble cell extracts. (C ) GST-b-catenin, or the GST-coupled cytoplasmic domain of VE-cadherin (GST-VE-cadcyto) was immobilized
to glutathione-S-sepharose beads and incubated with Triton X-100-soluble HUVEC cell lysates and probed by western blotting as indicated. Total cell lysates served as a positive control. *P , 0.05; **P , 0.005; ***P , 0.0005.
R. Kronstein et al.
of GST-b-catenin and the GST-coupled cytoplasmic domain of VE-cadherin (GST-VE-cadcyto) to precipitate caveolin-1 from
endo-thelial cell extracts. Indeed, GST-b-catenin, but not GST-VE-cadcyto
or GST-glutathione-sepharose, pulled down caveolin-1 (Figure 1C). These data indicate that caveolin-1 associates with the VE-cadherin/catenin complex in endothelial cells and requires b-catenin for interaction. Since PECAM-1 was shown to be associ-ated with both av/b3 integrin27 and b-catenin,28 we expected
caveolin-1 in PECAM-1 immunoprecipitates. Indeed, IP of PECAM-1 yielded a fraction containing both b-catenin and caveolin-1 (Figure 1A and B).
3.2 Recruitment of caveolin-1 to cell
junctions depends on the VE-cadherin/
catenin complex
Immunostaining of confluent cultures of HUVEC showed that a frac-tion of caveolin-1 localized along the cell juncfrac-tions and appeared superimposed with VE-cadherin, catenins, and PECAM-1 (Supplemen-tary material online, SI-1).
To determine whether VE-cadherin and/or the catenins, in particu-lar b-catenin, are important for the junction localization of caveolin-1, we analysed the distribution of caveolin-1 in two different cell systems: CHO cells and b-catenin-knockout endothelioma cells.21 CHO cells endogenously express caveolin-1 and small amounts of b-catenin, but not a-catenin, g-catenin, PECAM-1, VE-cadherin, or other classical cadherins (Figure 2A and B; Supplementary material online, SI-2). In wild-type CHO cells neither b-catenin nor caveolin-1 was localized to the cell junctions (Figure2A, a – d), while expression of VE-cadherin recruited both the bulk of the b-catenin and a fraction of the caveolin-1 (Figure 2A,e – h). Since VE-cadherin expression also induced up-regulation of both b-catenin and a-catenin,29 but not caveolin-1 (Figure 2B), we further investigated whether caveolin-1 recruitment to cell junctions was simply related to the increased amount of b-catenin in VE-cadherin-expressing CHO cells. When overexpressed in CHO cells (Figure2C), b-catenin was predominantly localized in the cytosol and caveolin-1 was not seen at cell junctions (Figure 2A; i – l). Consistently, we found by IP that association of caveolin-1 with b-catenin was high in VE-cadherin-expressing CHO cells, much lower in b-catenin-overexpressing CHO cells, and negli-gible in wild-type CHO cells (Figure2D). Since caveolin-1 is a com-ponent of a number of subcellular domains its interaction with b-catenin requires spatial proximity at junctions as seen in VE-cadherin-expressing CHO (Figure 2A; e – l) and endothelial cells (Figures3; Supplementary material online, SI-1). In addition, we used b-catenin-knockout endothelial cells21 that still expresses VE-cadherin, a- and g-catenin, and caveolin-1 (Supplementary material online, SI-3). Since immunoprecipiates of g-catenin also yield caveolin-1 (Figure1), we found a small fraction of caveolin-1 at cell junctions in b-catenin-knockout endothelial cells, as expected (Figure3A). Re-expression of b-catenin in b-catenin-knockout endo-thelial cells increased the junction localization of caveolin-1 by a factor of 2.5 (Figure3A and B). This amount of caveolin-1 is compar-able with the amount of caveolin-1 found in wild-type endothelial cells of both murine and human origin. Thus, the VE-cadherin/catenin complex is able to recruit caveolin-1 to cell junctions and b-catenin greatly facilitates this effect.
3.3 Thrombin stimulation increases
formation of caveolin-1 complexes
with b- and g-catenin and decreases
association with VE-cadherin
The functional relevance of caveolin-1 complex formation with VE-cadherin and catenins was tested by stimulation with thrombin, a serine protease that cleaves the PAR-1 receptor and consequently increases paraendothelial permeability.30 Impedance spectroscopy was used to determine TER.5 One unit of thrombin was found to transiently down-regulate TER to 70% of baseline values; the most sig-nificant effect in HUVEC was seen 10 – 25 min post-stimulation, with full recovery after 1 to 2 h (Figure4A). Furthermore, standardized IP of junction components revealed that thrombin stimulation decreased the association of caveolin-1 with VE-cadherin to about one-third after 20 min; this returned to pre-stimulation levels within 120 min (Figure 4B and C ). In contrast, thrombin stimulation increased caveolin-1 association with both b-catenin- and g-catenin by two-fold in Triton X-100-soluble cell extracts as shown by IP of the respective catenins (Figure4B, D, and E). We performed IPs of cate-nins instead of caveolin-1 as catecate-nins appear to be target proteins of caveolin-1 at cell junctions. Consistently, a fraction of b-catenin dis-sociated from VE-cadherin after thrombin stimulation, as assayed by VE-cadherin IP followed by b-catenin detection in Triton X-100 cell extracts. This effect became increasingly apparent when the b-catenin/VE-cadherin ratio was calculated (Figure4F and G). These effects were highly specific, since the amount of caveolin-1 bound to a-catenin and PECAM-1 remained unchanged in response to thrombin stimulation (Figure4B, H, and I ).
3.4 Remodelling of the VE-cadherin/catenin
complex during thrombin-induced opening
and reconstitution of cell junctions
Consistent with the thrombin-induced TER-minimum VE-cadherin (Figure5A), b-catenin (Figure 5B and C ), and g-catenin (not shown) underwent remodelling from a continuous to an interrupted and rope ladder-like pattern (Figure5). A fraction of caveolin-1 localized at the cell junctions under control conditions (Figure 5A) and underwent remodelling in accordance with the redistribution of the VE-cadherin/ catenin complex into a rope ladder-like pattern, as shown for b-catenin and caveolin-1 (Figure5C). This remodelling was accompanied by the characteristic loss of junction-associated actin filaments and the development of stress fibres (Supplementary material online, SI-4). During TER-recovery both caveolin-1 and the VE-cadherin/catenin complex reappeared at the junctions after 120 min, as shown for VE-cadherin and b-catenin (Figure5A and D). The thrombin-induced junction reorganization was also seen after stimulation of endothelial cells with the PAR-1-specific PAR-1-activated peptide (PAR-1AP; not shown). These data further support the concept that the impact of caveolin-1 on junction regulation occurs via interaction with the VE-cadherin/catenin complex.
3.5 Caveolin-1 is critical for the
transient breakdown of paracellular barrier
function
Since caveolin-1 can be activated by phosphorylation at tyrosine 14,31 we found a transient increase in phosphocaveolin-1 in b-catenin IP after thrombin stimulation (Figure 6A), whereas the amount of
Figure 2 VE-cadherin/catenin complexes target caveolin-1 to cell junctions in CHO cells. Wild-type CHO cells (CHO wt), CHO cells expressing VE-cadherin (CHO-VE-cad), and CHO cells overexpressing b-catenin (CHO+ b-cat) were immunolabelled for (A) VE-cadherin (panels a, e and i) or triple-labelled for b-catenin, caveolin-1, and nuclear DNA (DAPI). Yellow colour indicates co-localization of b-catenin and caveolin-1 (arrows). Expression of VE-cadherin but not b-catenin increased recruitment of both b-catenin and caveolin-1 to cell junctions. Inserts indicate confocal sec-tions in the vertical plane of the z-stacks taken at the position shown by the white lines. Stars indicate same cell in the series. Bars ¼ 10 mm. (B) Total cell lysates (20 mg/lane) were analysed by western blotting as indicated. HUVEC, positive control. Expression of VE-cadherin CHO wild-type (wt) increased expression of b- and a-catenin. Alpha-tubulin, internal control. (C ) Total cell lysates (10 mg/lane) as indicated were probed by western blotting for b-catenin. Increased degradation of b-catenin is seen only in CHO wt+ b-cat cells. (D) IP of b-catenin from cell lysates as indicated was probed for either VE-cadherin or caveolin-1.
R. Kronstein et al.
phosphorylated caveolin-1 precipitated with the VE-cadherin antibody remained unchanged (Figure6A). To prove this concept, we generated caveolin-1-knockout endothelioma cells and expressed the wild-type caveolin-1 or caveolin-1 14Y/F (cav-1-Y/F), a mutant form of caveolin-1 that can no longer be phosphorylated at tyrosine 14 (Figure6B). The caveolin-1-knockout endothelioma cells are shown in the Supplemen-tary material online, SI-5. Thrombin stimulation of caveolin-12/2 endothelial cells (cav2/2cells) did not decrease the TER, but rather increased it slightly (Figure 6C). Expression of caveolin-1 in caveolin-12/2 cells (cav2/2cells+ cav-1) completely reconstituted the thrombin response, while expression of the caveolin-1 Y/F mutant (cav2/2cells+ cav-1Y/F) did not. Thus, cav2/2cells + cav-1Y/F behaved like cav2/2cells lacking caveolin-1 (Figure6C). Under these conditions, the association of caveolin-1 with b-catenin increased in cav2/2cells+ cav-1 but not in cav2/2cells + cav-1Y/F (Figure 6D
and E). Consistent with the hypothesis that phosphocaveolin-1 med-iates dissociation of b-catenin from VE-cadherin, we found decreased amounts of b-catenin in VE-cadherin IPs after thrombin application in cav2/2cells+ cav-1 but not in cav2/2cells or cav2/2cells + cav-1Y/F. Rather, the addition of thrombin to cav2/2cells and cav2/2 cells+ cav-1Y/F increased formation of b-catenin complexes with VE-cadherin (Figure6F). Together, these data provide direct evidence that phosphocaveolin-1 is required to open intercellular junctions in response to thrombin by increasing the formation of complexes with catenins that dissociate from VE-cadherin.
4. Discussion
Regulation of paracellular endothelial barrier function requires a concerted interplay between cell adhesion molecules and the Figure 3 Caveolin-1 is increasingly recruited to endothelial cell junctions in the presence of b-catenin. (A) Immunolabelling of b-catenin, g-catenin, caveolin-1, and nuclear DNA (DAPI) in b-catenin-deficient endothelial cells (b-cat 2/2) and b-cat 2/2 re-expressing b-catenin (b-cat 2/2+ b-cat) as indicated. Boxed areas within panels indicate those areas also shown at higher magnification. Z-stack: confocal sections in the vertical plane of the z-stacks taken at the position of the white lines. Expression of b-catenin increased junction localization of caveolin-1, as seen by yellow staining, indicating co-localization of caveolin-1 and g-catenin. Bars ¼ 10 mm. (B) Quantitative analysis of the co-localized caveolin-1 and g-catenin at cell junctions. At least 200 cells of each type were analysed.
Figure 4 Thrombin-induced barrier function decrease is associated with increased formation of catenin/caveolin-1 complexes that dissociate from VE-cadherin. (A) Transendothelial electrical resistance (TER) was determined by impedance spectroscopy and plotted relative to baseline TER values {TER/TER (0)} in HUVEC with and without 1 U/mL of thrombin. TER(0) values ranged between 8 and 15 Vcm2
. n ¼ 3. ***P , 0.0005. (B – I) HUVEC were treated with 1 U/mL of thrombin followed by IP of adherens junction proteins and subsequent quantitative western blotting, as indicated. Prior to IP, nuclei-free Triton X-100 cell extracts from each experimental series were adjusted for equal protein content and equal volume. A sample of 20 mL was blotted and probed for a-tubulin content for internal control and quantification of relative protein amounts of co-precipitated proteins (panels C1 – E1, G1 – I1). (B) Densitometry of western blots after IP with antibodies as indicated, n ¼ 3. (C2E, H – I ) One out of the three independent western blots that were used for quantification as indicated is shown. (C ) Caveolin-1 dissociated transiently from the VE-cadherin/catenin complex after 20 min of thrombin stimulation, while caveolin-1 transiently increased its association with both (D and E) b-catenin and g-catenin. (F and G) b-Catenin/VE-cadherin ratios were determined by densitometry of western blots (one out of three is seen in G) after thrombin stimulation. (H and I ) Association of caveolin-1 with a-catenin and PECAM-1 remained unchanged. *P , 0.05; **P , 0.005; ***P , 0.0005.
R. Kronstein et al.
junction-associated actin cytoskeleton. The ability of caveolin-1 to bind to and control certain regulatory molecules, such as src kinases, G-proteins, and eNOS, as well as its ability to modulate calcium signalling and permeability6,8,19,32,13 indicates a critical role in junction dynamics and integrity. Here, we aimed to find novel clues as to how caveolin-1 controls cell adhesion and barrier function in endothelial cells. To this end, we demonstrated a caveolin-1-controlled mechanism that is directly linked to VE-cadherin-mediated cell adhesion in response to thrombin stimulation.
The most significant findings of this study are (i) that caveolin-1 associates with and is targeted by the VE-cadherin/catenin complex to endothelial cell junctions and (ii) that phosphocaveolin-1 is critical for the opening of endothelial cell junctions in response to thrombin stimulation. Association of catenins with caveolin-1 was demonstrated by IPs, GST-pull-down assays (Figure1), and expression studies using CHO cells (Figure2) and a b-catenin-knockout cell line (Figure3). The data are in line with other reports showing that Vcadherin and E-cadherin are associated with caveolin-113,15 as well as the identifi-cation in zebrafish of a binding motif that mediates the interaction
of caveolin-1 with b-catenin.16Thus, a direct interaction of mamma-lian caveolin-1 with mammamamma-lian b-catenin might occur, but this has to be tested in the future.
The functional relevance of our findings was further demonstrated in primary cultures of HUVEC and a caveolin-1-knockout endothelial cell line after stimulation with thrombin. Certain signalling pathways become activated by thrombin through G-protein-coupled PAR-1. This includes activation of Rho-GTPases, non-receptor tyrosine kinases such as Pyk2 and src-kinase, myosin light chain kinases and phosphatases,33,34 and actin/myosin filament sliding mechanisms.35,36 Furthermore, caveolin-1 becomes tyrosine phosphorylated by src-kinases,7a process that has also been described to be critical in barrier function regulation.34Indeed, thrombin induces tyrosine phos-phorylation of caveolin-1 by hitherto unknown non-receptor tyrosine kinases, which in turn increases complex formation with both b-catenin and g-catenin and weakens binding to VE-cadherin. This is reflected in the decreased amounts of both b-catenin and caveolin-1 that were precipitated from soluble Triton X-100 cell extracts with VE-cadherin and the increased amounts of caveolin-1 that were Figure 5 Thrombin-induced reorganization of junction molecules. Immunolabelling of HUVEC after treatment with either 1 U/mL thrombin or buffer as indicated. (A – D) Thrombin stimulation induced a transient redistribution of VE-cadherin, b-catenin, and caveolin-1 into a rope ladder-like pattern within 15 min that recovered after 120 min. Boxes indicate areas of higher magnification. Arrows point to cell junctions. Under control con-ditions (B) a fraction of caveolin-1 co-localized at cell junctions and with a fraction of b-catenin (yellow staining, arrowheads). Thrombin administration dissociated caveolin-1 from cell junctions and (C ) colocalized with rope ladder-like b-catenin staining (arrows). (D) Within 120 min both caveolin-1 and b-catenin reappeared at junctions as under control conditions. Panels on the right-hand side show magnification of the respective boxes. Nuclei are stained with DAPI. Scale bars ¼ 20 mm.
Figure 6 Phosphocaveolin-1 is required to open cell junctions in response to thrombin. (A) Quantitative western blot of phosphocaveolin-1 using a Y-14 phosphocaveolin-1-specific antibody after IP of adherens junction molecules from HUVEC treated with thrombin as indicated. Phosphocaveolin-1 was transiently increased in the b-catenin IP, whereas the amount of phosphocaveolin-1 that associated with VE-cadherin remained unchanged. Pervanadate (PV) served as a positive control. *P , 0.05; **P , 0.005; ***P , 0.0005. (B) Expression of caveolin-1 and the 14 Y/F-caveolin-1-mutant in cav-12/2 endothelial cells by lentiviral gene transfer. Alpha-tubulin served as internal protein loading control. (C ) TER of endothelioma cells as indicated in response to thrombin. Only caveolin-1-expressing cells showed the characteristic decrease in TER. (D) Quantitative western blotting of b-catenin IP from endothelioma cells as indicated. (E) Immunoprecipitates of b-catenin from indicated cell lines were probed for caveolin-1 after thrombin stimulation. (E1) Western blot of caveolin-1 from total Triton X-100 cell extracts as indicated. a-tubulin served as a further loading control. (F ) IP of VE-cadherin from indicated cell lines was probed for b-catenin after thrombin stimulation. One representative blot from three independent experiments is shown. Quantitative analysis of three independent experiments is shown. *P , 0.05; **P , 0.005.
R. Kronstein et al.
precipitated with b- and g-catenin. The non-ionic detergent Triton X-100 dissociates protein complexes only when the association is weak,37 and it is reasonable to assume that the b-catenin/-phosphocaveolin-1 complex is dissociated from VE-cadherin after Triton X-100 extraction. These results are also consistent with the localization of caveolin-1 and b-catenin in a rope ladder-like pattern after thrombin stimulation (Figure 5C). While histamine stimulation has been shown to increase the association of b-catenin with PECAM-1,38 a mechanism that might also reduce the association of b-catenin with VE-cadherin, the interaction of PECAM-1 with caveolin-1 seems to be of less importance after thrombin stimulation (Figure4B and I ). Another possible role of caveolin-1 in the regulation of endothelial barrier function was recently reported in the form of hydrogen peroxide-induced dissociation of b-catenin and caveolin-1 in rat lung microvascular endothelial cells.13 Hydrogen peroxide is an important but also an unspecific stimulus that leads to increased endothelial permeability, and a different mechanism utilizing caveolin-1- seems to be involved. This shows the existence of differ-ent mechanisms that seem to depend on the respective stimuli but might also depend on extraction buffers and cell type. However, both studies13(and the present study) point to the critical importance of caveolin-1 in barrier function regulation.
Catenins occupy a strategically important position at the adherens junctions since they link VE-cadherin to the cytoskeleton. Weakening of this interaction by phosphocaveolin-1, together with Rho activation in response to thrombin stimulation, might facilitate uncoupling of the VE-cadherin/catenin complex from the junction-associated actin cytoskeleton, leading to the formation of stress fibres and the devel-opment of contractile forces.35In addition, caveolin-1 interacts with the actin filament cross-linking protein filamin,39and this interaction might be important in uncoupling the VE-cadherin/catenin complex from the actin cytoskeleton. Previous work showed lateral clustering of the VE-cadherin/catenin complex and a rac-1-dependent recruit-ment of actin filarecruit-ments to adherens junctions in response to physio-logical shear stress, a stimulus that tightens endothelial cell junctions.5 Thus, thrombin-induced weakening of b- and g-catenin binding to VE-cadherin might also lead to reduced lateral clustering of the VE-cadherin/catenin complex. However, whether and to what extent lateral clustering contributes to the regulation of the VE-cadherin/catenin complex and its connection to the actin cytoske-leton during inflammation has yet to be determined.
Here, we show that PAR-1 activation leads to phosphorylation of caveolin-1, which in turn is essential to open endothelial cell junctions but is independent of the signalling that leads to ERK1/2 phosphoryl-ation.40Thus, PAR-1-mediated activation of caveolin-1 turns on mul-tiple pathways that appear to be independent of each other. The data are in line with b-catenin-mediated, thrombin-induced long-term stimulation of gene expression,41 and support the concept that b- and g-catenin are targets in the opening of cell junctions. Other stimuli such as histamine also target b-catenin,42 and g-catenin was shown to be targeted by VE phosphotyrosine phosphatase during leu-cocyte diapedesis.43We assume that caveolin-1 plays a role in these processes as well, but this has to be investigated in a separate study. The role of caveolin-1 in the regulation of endothelial barrier func-tion has also been studied using a siRNA approach. Deplefunc-tion of caveolin-1 increases constitutive endothelial permeability and reduces VE-cadherin and b-catenin levels12(Supplementary material online, SI-5). However, the caveolin-1-knockout cells still formed a barrier (Supplementary material online, SI-5) and responded to
thrombin with a slight increase in barrier function, as seen by an increase in TER. This phenomenon demonstrated a counter-regulatory mechanism that is activated in parallel by thrombin but is independent of caveolin-1. In contrast, caveolin-1 is essential in the opening of cell junctions in response to thrombin. Thrombin increases paracellular permeability in a process requiring the phosphorylation of caveolin-1 at tyrosine 14.
Supplementary material
Supplementary material is available at Cardiovascular Research online.
Acknowledgements
We thank Christine Mund, Antje Messer, and Annelie Ahle for excellent technical assistance. Teymuras Kurzchalia kindly provided us with breeding pairs of caveolin-1 knockout mice. William Weis and Sabine Pokutta (Stanford, USA) generously provided the plasmids for recom-binant GST-b-catenin, and Dietmar Vestweber (Mu¨nster, Germany) provided the plasmids for the generation of recombinant GST-VE-cadherin cytoplasmic domain and the VE-cadherin-expressing CHO cells. Elisabetta Dejana kindly provided the mouse b-catenin endothelioma cells and Rolf Kemler (Tu¨bingen, Germany) the mouse b-catenin cDNA. We are grateful to Michele Solimena, Kai Simons, and Rolf Jessberger for helpful discussion.
Conflict of interest: none declared.
Funding
This work was supported by grants from the Deutsche Forschungsge-meinschaft (SCHN 430/3 – 3 and SCHN 430/6 – 1) and the CRTD Dresden.
References
1. Dejana E, Tournier-Lasserve E, Weinstein BM. The control of vascular integrity by endothelial cell junctions: molecular basis and pathological implications. Dev Cell 2009;16:209 – 221.
2. Vestweber D. VE-cadherin: the major endothelial adhesion molecule controlling cel-lular junctions and blood vessel formation. Arterioscler Thromb Vasc Biol 2008;28: 223 – 232.
3. Pokutta S, Weis WI. Structure and mechanism of cadherins and catenins in cell-cell contacts. Annu Rev Cell Dev Biol 2007;23:237 – 261.
4. Abe K, Takeichi M. EPLIN mediates linkage of the cadherin catenin complex to F-actin and stabilizes the circumferential actin belt. Proc Natl Acad Sci U S A 2008;105:13 – 19. 5. Seebach J, Donnert G, Kronstein R, Werth S, Wojciak-Stothard B, Falzarano D et al. Regulation of endothelial barrier function during flow-induced conversion to an arter-ial phenotype. Cardiovasc Res 2007;75:596 – 607.
6. Liu P, Rudick M, Anderson RG. Multiple functions of caveolin-1. J Biol Chem 2002;277: 41295 – 41298.
7. Parton RG, Simons K. The multiple faces of caveolae. Nat Rev Mol Cell Biol 2007;8: 185 – 194.
8. Bauer PM, Yu J, Chen Y, Hickey R, Bernatchez PN, Looft-Wilson R et al. Endothelial-specific expression of caveolin-1 impairs microvascular permeability and angiogenesis. Proc Natl Acad Sci U S A 2005;102:204 – 209.
9. Schubert W, Frank PG, Razani B, Park DS, Chow CW, Lisanti MP. Caveolae-deficient endothelial cells show defects in the uptake and transport of albumin in vivo. J Biol Chem 2001;276:48619 – 48622.
10. Woodman SE, Ashton AW, Schubert W, Lee H, Williams TM, Medina FA et al. Caveolin-1 knockout mice show an impaired angiogenic response to exogenous stimuli. Am J Pathol 2003;162:2059 – 2068.
11. Millan J, Hewlett L, Glyn M, Toomre D, Clark P, Ridley AJ. Lymphocyte transcellular migration occurs through recruitment of endothelial ICAM-1 to caveola- and F-actin-rich domains. Nat Cell Biol 2006;8:113 – 123.
12. Song L, Ge S, Pachter JS. Caveolin-1 regulates expression of junction-associated pro-teins in brain microvascular endothelial cells. Blood 2007;109:1515 – 1523. 13. Sun Y, Hu G, Zhang X, Minshall RD. Phosphorylation of caveolin-1 regulates
oxidant-induced pulmonary vascular permeability via paracellular and transcellular pathways. Circ Res 2009;105:676 – 685.
14. Chang SH, Feng D, Nagy JA, Sciuto TE, Dvorak AM, Dvorak HF. Vascular permeability and pathological angiogenesis in caveolin-1-null mice. Am J Pathol 2009;175: 1768 – 1776.
15. Galbiati F, Volonte D, Brown AM, Weinstein DE, Ben-Ze’ev A, Pestell RG et al. Caveolin-1 expression inhibits Wnt/catenin/Lef-1 signaling by recruiting beta-catenin to caveolae membrane domains. J Biol Chem 2000;275:23368 – 23377. 16. Mo S, Wang L, Li Q, Li J, Li Y, Thannickal VJ et al. Caveolin-1 regulates dorsoventral
patterning through direct interaction with beta-catenin in zebrafish. Dev Biol 2010; 344:210 – 223.
17. Schnittler HJ, Puschel B, Drenckhahn D. Role of cadherins and plakoglobin in inter-endothelial adhesion under resting conditions and shear stress. Am J Physiol 1997; 273:H2396 – H2405.
18. Nawroth R, Poell G, Ranft A, Kloep S, Samulowitz U, Fachinger G et al. VE-PTP and VE-cadherin ectodomains interact to facilitate regulation of phosphorylation and cell contacts. EMBO J 2002;21:4885 – 4895.
19. Drab M, Verkade P, Elger M, Kasper M, Lohn M, Lauterbach B et al. Loss of caveolae, vascular dysfunction, and pulmonary defects in caveolin-1 gene-disrupted mice. Science 2001;293:2449 – 2452.
20. Reiss Y, Hoch G, Deutsch U, Engelhardt B. T cell interaction with ICAM-1-deficient endothelium in vitro: essential role for ICAM-1 and ICAM-2 in transendothelial migration of T cells. Eur J Immunol 1998;28:3086 – 3099.
21. Cattelino A, Liebner S, Gallini R, Zanetti A, Balconi G, Corsi A et al. The conditional inactivation of the beta-catenin gene in endothelial cells causes a defective vascular pattern and increased vascular fragility. J Cell Biol 2003;162:1111 – 1122.
22. Pokutta S, Weis WI. Structure of the dimerization and beta-catenin-binding region of alpha- catenin. Mol Cell 2000;5:533 – 543.
23. Schwencke C, Schmeisser A, Walter C, Wachter R, Pannach S, Weck B et al. Decreased caveolin-1 in atheroma: loss of antiproliferative control of vascular smooth muscle cells in atherosclerosis. Cardiovasc Res 2005;68:128 – 135. 24. Butz S, Stappert J, Weissig H, Kemler R. Plakoglobin and beta-catenin: distinct but
closely related letter]. Science 1992;257:1142 – 1144.
25. Naldini L, Blomer U, Gallay P, Ory D, Mulligan R, Gage FH et al. In vivo gene delivery and stable transduction of nondividing cells by a lentiviral vector. Science 1996;272: 263 – 267.
26. Dieckmann-Schuppert A, Schnittler HJ. A simple assay for quantification of protein in tissue sections, cell cultures, and cell homogenates, and of protein immobilized on solid surfaces. Cell Tissue Res 1997;288:119 – 126.
27. Wong CW, Wiedle G, Ballestrem C, Wehrle-Haller B, Etteldorf S, Bruckner M et al. PECAM-1/CD31 trans-homophilic binding at the intercellular junctions is indepen-dent of its cytoplasmic domain; evidence for heterophilic interaction with integrin alphavbeta3 in Cis In Process Citation]. Mol Biol Cell 2000;11:3109 – 3121. 28. Ilan N, Cheung L, Pinter E, Madri JA. Platelet-endothelial cell adhesion molecule-1
(CD31), a scaffolding molecule for selected catenin family members whose binding is mediated by different tyrosine and serine/threonine phosphorylation. J Biol Chem 2000;275:21435 – 21443.
29. Breviario F, Caveda L, Corada M, Martin-Padura I, Navarro P, Golay J et al. Functional properties of human vascular endothelial cadherin (7B4/cadherin-5), an endothelium-specific cadherin. Arterioscler Thromb Vasc Biol 1995;15:1229 – 1239.
30. Bogatcheva NV, Garcia JG, Verin AD. Molecular mechanisms of thrombin-induced endothelial cell permeability. Biochemistry (Mosc) 2002;67:75 – 84.
31. Cao H, Courchesne WE, Mastick CC. A phosphotyrosine-dependent protein inter-action screen reveals a role for phosphorylation of caveolin-1 on tyrosine 14 recruit-ment of C-terminal Src kinase. J Biol Chem 2002;277:8771 – 8774.
32. Li S, Couet J, Lisanti MP. Src tyrosine kinases, Galpha subunits, and H-Ras share a common membrane-anchored scaffolding protein, caveolin. Caveolin binding nega-tively regulates the auto-activation of Src tyrosine kinases. J Biol Chem 1996;271: 29182 – 29190.
33. Beckers CM, van Hinsbergh VW, van Nieuw Amerongen GP. Driving Rho GTPase activity in endothelial cells regulates barrier integrity. Thromb Haemost 2010;103: 40 – 55.
34. Hu G, Place AT, Minshall RD. Regulation of endothelial permeability by Src kinase sig-naling: Vascular leakage versus transcellular transport of drugs and macromolecules. Chem Biol Interact 2008;171:177 – 89.
35. Schnittler HJ, Wilke A, Gress T, Suttorp N, Drenckhahn D. Role of actin and myosin in the control of paracellular permeability in pig, rat and human vascular endothelium. J Physiol (Lond) 1990;431:379 – 401.
36. Wysolmerski RB, Lagunoff D. Involvement of myosin light-chain kinase in endothelial cell retraction. Proc Natl Acad Sci U S A 1990;87:16 – 20.
37. Patton WF, Dhanak MR, Jacobson BS. Differential partitioning of plasma membrane proteins into the triton X-100-insoluble cytoskeleton fraction during concanavalin A-induced receptor redistribution. J Cell Sci 1989;92(Pt 1):85 – 91.
38. Biswas P, Canosa S, Schoenfeld D, Schoenfeld J, Li P, Cheas LC et al. PECAM-1 affects GSK-3beta-mediated beta-catenin phosphorylation and degradation. Am J Pathol 2006; 169:314 – 324.
39. Stahlhut M, van Deurs B. Identification of filamin as a novel ligand for caveolin-1: evi-dence for the organization of caveolin-1-associated membrane domains by the actin cytoskeleton. Mol Biol Cell 2000;11:325 – 337.
40. Russo A, Soh UJ, Paing MM, Arora P, Trejo J. Caveolae are required for protease-selective signaling by protease-activated receptor-1. Proc Natl Acad Sci U S A 2009; 106:6393 – 6397.
41. Beckers CM, Garcia-Vallejo JJ, van Hinsbergh VW, van Nieuw Amerongen GP. Nuclear targeting of beta-catenin and p120ctn during thrombin-induced endothelial barrier dysfunction. Cardiovasc Res 2008;79:679 – 688.
42. Guo M, Breslin JW, Wu MH, Gottardi CJ, Yuan SY. VE-cadherin and beta-catenin binding dynamics during histamine-induced endothelial hyperpermeability. Am J Physiol Cell Physiol 2008;294:C977 – C984.
43. Nottebaum AF, Cagna G, Winderlich M, Gamp AC, Linnepe R, Polaschegg C et al. VE-PTP maintains the endothelial barrier via plakoglobin and becomes dissociated from VE-cadherin by leukocytes and by VEGF. J Exp Med 2008;205:2929 – 2945.
R. Kronstein et al.
